机器灵巧手基于触滑觉信息协同的自适应力控制方法研究

项目名称： 机器灵巧手基于触滑觉信息协同的自适应力控制方法研究

项目编号： No.61503123

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 吴兰

作者单位： 河南工业大学

项目金额： 22万元

中文摘要： 随着社会的进步与人们生活水平的不断提高，对各类机器人的需求越来越多。特别是一些特殊行业对机器灵巧手精确抓取的要求越来越高。由于现有灵巧手抓取方法未考虑基于目标对象特征的初始抓取力分配和抓取过程中基于触觉与滑觉联动关系的多指力再分配方案等，常会造成对抓取物体不同程度的损伤。为此，本项目将以一类非刚体物体为抓取对象，建立一套机器灵巧手基于触滑觉信息协同的自适应力控制方法，重点开展以下研究：未知环境下抓取对象特征变量的多级层次化建模；抓取过程中抓取力的动态建模方法；基于触滑觉信息协同的抓取力自适应控制策略；并基于仿真和实际数据开展相应模型和控制算法的验证与测试。重点解决在建立上述方法过程中存在的一些关键科学问题：抓取力与触觉和滑觉信息之间联动关系的建立；多指手随滑觉和触觉信息抓取力的再分配策略。取得一些创新性研究成果，为提高灵巧手在实际应用中的精确抓取水平，打下坚实的理论基础和关键技术上的储备。

中文关键词： 数据驱动建模；多层次建模；混合建模

英文摘要： With the progress of the society and the continuous improvement of people's living standard, more and more all kinds of robots are needed. Especially for special industries, the request of dexterous hand machine accurately grasping is higher and higher. Due to the existing dexterous hand grasping methods haven’t considering the initial grasping force distribution based on the characteristics of the target object and based on the tactile and slip sensation in the process of fetching more relationship refers to force redistribution scheme, etc., it will often cause varying degrees of damages to the grab objects. Therefore, this project will use a kind of non-rigid objects as grab objects, establishes a set of machine dexterous hand based on the touch slippery sense of information synergy self-adaptive control method, and focus on the multilevel hierarchical modeling of grab objects characteristic variables under unknown environment; Dynamic modeling method of grasping force in the process of fetching;Grasping force self-adaptive control strategy based on touch and slippery sense information synergy; And carry out the verification and testing of corresponding model and control algorithm based on the simulation and real data. The key point is to solve some of the key scientific problems in the process of building the solution which are the building of the follow-up relationship between the grasping force and tactile and slip sensation information ; The redistribution policies of grasping force of hand with more than one finger changes over over slippery and touch information . Achieve some innovative research results, lay a solid theoretical foundation and the key technical reserves to improve the level of dexterous hand in the actual application of precise grab.

英文关键词： data-driven modeling;multi-layer modelling;hybrid modeling